Hi, I'm Tarit Witworrasakul
I'm a second-year Computer Science student at the University of British Columbia (UBC) with a passion for solving problems and building solutions to make life easier through programming. My journey began with Arduino robot programming, which sparked my interest in creating innovative systems. Over time, I expanded my skill set to include web development, mobile apps, and embedded systems.
Work Experience
- Conducted an analysis of the sales team's needs for a customer identification solution, defining essential functionalities and designing user flows in collaboration with a UX/UI designer to develop a Figma prototype
- Developed a real-time facial recognition web application using Next.js, featuring a user interface for customer identification with 95% accuracy via webcam, along with an admin page for managing customer data
- Built a backend service in Python with FastAPI for face vector comparison using ChromaDB, integrated with a PostgreSQL database for storing User Data, and linked it to the existing internal customer database
- Implemented a robust authentication system featuring role-based access control with NextAuth, utilizing JWT sessions and email OTP for secure login, and employed Next.js middleware for token verification
- Deployed the application and databases to an on-premises Ubuntu server using Docker and Docker Compose
Technical Extracurriculars
- Contributed to the development of autonomous drones competing in the the Aerial Evolution Association of Canada 2024 competition, achieving 2nd place in the proposals phase and 3rd place in the demonstration phase
- Successfully migrated 15+ endpoints in MissionPlanner-Scripts, a REST APIs and WebSocket interface between a backend system and ArduPilot, to a cross-platform architecture using Pymavlink and Flask, eliminating reliance on Windows-only MissionPlanner client scripts
- Crafted OpenAPI Specifications and Postman collections to facilitate team collaboration and API integration
- Developed a pipeline for drone target detection using a fine-tuned YOLOv8 model with tiled inference
- Created dataset generator program using OpenCV and imgaug, eliminating the need for manual data labeling
Projects
An app that helps individuals with Autism Spectrum Disorder (ASD) recognize sarcasm in conversations—making social interactions smoother and reducing misunderstandings
Finetuning No Language Left Behind model for Thai to English machine translation using SCB-1M and OPUS dataset
Mobile application to index images based on the content of the image by using object detection and optical character recognition
Android application for controlling NASA's free-flying robot (Astrobee) in the Japanese Experimental Module “Kibo” of the International Space Station (ISS)
Logistics robot that travels on a simulated moon surface, detects radioactive substances and moves them into storage boxes
Logistics robot that drives on the road lane and transports objects to the correct destination by reading the QR code